LangChain Reportedly Raises $300M at $3B to Take On Model Makers' Agent Tools
CapitalG led the round, the Wall Street Journal reported, at 2.4 times LangChain's October 2025 price. LangChain plans to nearly double headcount to 850 as OpenAI, Anthropic and Google build their own agent tools.
TL;DR
- LangChain raised $300 million in a CapitalG-led round at a $3 billion valuation, 2.4 times its October 2025 price, the Wall Street Journal reported.
- LangChain sells LangSmith as an agent platform that works across models, while OpenAI, Anthropic and Google build agent tools around their own.
- LangChain plans to grow from 435 employees to 850 by the end of 2027. OpenAI's Agent Builder, one of the lab-built agent tools, shuts down on November 30.
LangChain, which makes open-source tools for building AI agents and sells a platform to run them, more than doubled its valuation to $3 billion, the Wall Street Journal reported on November 12.
CapitalG led a $300 million round, FourWeekMBA reported, citing the Journal. Dealroom's summary of the report lists IVP and Sapphire as returning investors and GIC and DFJ Growth as new ones. The price is 2.4 times the $1.25 billion LangChain set with its Series B in October 2025. LangChain hadn't announced the round on its blog as of November 12.
The money backs LangSmith, LangChain's paid platform, as OpenAI, Anthropic and Google build agent tools around their own models.
What LangChain sells
Harrison Chase released LangChain as a side project in October 2022, while working as a machine learning engineer at Robust Intelligence, Contrary Research reports. ChatGPT launched the next month and brought the first wave of developers. Chase incorporated the company in 2023 with Ankush Gola, a former Robust Intelligence colleague, and Benchmark led a $10 million seed that April.
The open-source frameworks, LangChain and LangGraph, had a combined 90 million monthly downloads in October 2025, the company said. LangSmith is the paid layer for tracing, testing and deploying agents built with them. LangChain's pitch is that "agents are easy to prototype but hard to ship to production." Chase described the goal as helping large enterprises build, manage and deploy their own agents, according to Dealroom's summary. LangSmith's monthly trace volume grew 12 times year over year as of October 2025.
LangChain's case studies cite Workday and Snyk
The Series B post named Replit, Clay, Harvey, Rippling, Cloudflare, Workday and Cisco as customers. LangChain's case studies report these results:
- Workday cut the time to build a prototype agent from four or five months to under a day, and its agent platform team grew from four people to about 30.
- Snyk's support agent has handled more than 60,000 customer queries and resolves over 85% of sessions without a support ticket.
Alex Gatto, a senior machine learning engineer at Workday, said there was "no way that we could have delivered value to our customers without LangSmith's speed and maturity."
LangSmith's last revenue estimate is 17 months old
LangChain hasn't published a revenue figure. TechCrunch estimated LangSmith ARR at $12 million to $16 million as of June 2025, SiliconANGLE reported at the Series B. A LangChain spokesperson later told Fortune that revenue had grown.
How LangChain will spend the money
LangChain has 435 employees and plans to reach 850 by the end of 2027, according to Dealroom's summary.
CapitalG partner Mo Jomaa told the Journal that LangChain can work as an independent agent-development platform for enterprises, Dealroom reported. Jomaa also pointed to model makers trying to lock customers into their own tools.
Model makers are building the same agent layer
OpenAI, Anthropic, Amazon and Microsoft are building their own agent infrastructure, Dealroom's note on the round says. TechCrunch reported in July 2025 that model providers had added features to their own APIs that once set LangChain apart.
Those tools can also be retired. OpenAI told developers in June that Agent Builder, the visual tool it launched in October 2025, will shut down on November 30. It points them to its Agents SDK and ChatGPT Workspace Agents.
LangChain's answer is neutrality. "The harness lock-in is going to be harder to unwind than the model lock-in itself, because the harness is where your business logic lives," LangChain's Neil Dahlke wrote in a June post that named Anthropic's Claude Agent SDK, OpenAI's Agents API and Vertex AI Agent Builder. In November, LangChain's blog announced LangSmith running in customers' own accounts on AWS, Azure and Google Cloud.
LangChain is selling the freedom to switch models
Every lab in Dahlke's post sells agent tools built around its own models, and each has a reason to keep customers on them. Developers on OpenAI's Agent Builder have until November 30 to move off it.
no way that we could have delivered value to our customers without LangSmith's speed and maturity
LangChain plans to reach 850 employees by the end of next year, and OpenAI's Agent Builder, a rival agent tool, is scheduled to shut down on November 30.
